Evaluating ‘Suicide Surveillance Technologies’ at high-risk locations

10 September 2024 marks the 22nd annual World Suicide Prevention Day – designed to raise awareness, inspire action and promote change to help prevent suicide globally. This year's theme is 'changing the narrative' on suicide, encouraging openness in place of silence. Suicide is often the 'elephant in the room', something we are all aware of and sometimes even exposed to or personally affected by, but something that many people still feel unable to discuss until, sadly, it may be too late. This year it is time to 'start the conversation' around suicide and what can be done to prevent it.

Around a third of suicides in the United Kingdom take place in public locations, and this increased 'visibility' can mean there are greater opportunities to intervene and prevent death. It's time to consider the paradox of these suicides in public places – in the open, often in front of witnesses, but remaining stigmatised and taboo. 

Public places and their potential

The term 'public place' covers a broad range of location types from coastal locations to the railways, each presenting different means, environmental challenges, footfall, and even oversight. These are complex environments with unique features which may not always be interchangeable with one another.

We see reports in the media of passersby who have intervened in public places, sometimes even putting themselves in dangerous situations in order to save somebody from suicide. There are notorious photos documenting desperate attempts from the public and emergency services to bring someone back to safety, accompanied by stories of how they've stayed in touch and remained lifelong friends after sharing such a vulnerable moment. Anyone who has taken a train lately will likely have seen Samaritans posters on the platforms encouraging people to reach out if they are feeling distressed. Similarly, their 'small talk saves lives' campaign emphasises how simply making small talk with someone displaying worrying behaviours may be enough to interrupt their thought process and potentially save their life. This underscores ongoing efforts to break the silence surrounding suicide and encourage openness and empathy, i.e., changing the narrative.

In addition to opportunities for the general public to intervene, suicides in public places also highlight locations where measures can be implemented by local councils, governments, charities, community groups and landowners to prevent further deaths. These have been outlined in detail in a practice resource from Public Health England, and may include interventions such as 1) restricting access to dangerous locations by closing certain areas or installing barriers, 2) increasing the capacity for human intervention by training relevant staff or installing CCTV to monitor an area, 3) signage with contact details for mental health charities/ organisations or signposting to a place of safety, and 4) changing the public image of the location so it is not associated with being a suicide site. 

This article will focus on the second category, specifically the use of surveillance technologies for identifying a 'vulnerable' person and the consequent role they can play in suicide prevention.

'Smart' Surveillance Technologies

Smart Surveillance Technologies are those which go beyond simply capturing and recording an environment (e.g. CCTV cameras) to in some way enhance or automate the surveillance process. This might involve the use of sensors, for instance, to detect the presence of an individual in a monitored environment via body heat (e.g. passive infrared), proximity (e.g. radar), or vibrations (e.g. microphonic sensors). Artificial Intelligence (AI) can also be implemented, for instance, to detect specific behaviours (e.g. lingering), vehicles (i.e. via automated number plate recognition) or people (e.g. using facial recognition) of interest in camera footage. 

Some of these technologies are arguably not new phenomena, but the abilities and performance of smart surveillance technologies are evolving rapidly. As such they are often being implemented in public places around the world for crime prevention and, increasingly, for suicide prevention. 

Whilst the technologies by themselves are limited in what they can do to support someone in distress, if a system is able to identify someone at risk, it could potentially help to facilitate an intervention by alerting the relevant service(s). For example, one coastal location in Australia implemented CCTV and a 'virtual fence', alerting the local police whenever someone crosses a fence entering into an area of danger. Research suggests that officers felt the system was helpful for detecting people at risk, allowing a team to be sent to intervene. However, they also raised concerns about the potential impact of individuals being aware of it, such as changing their behaviour to avoid detection or going to a different location. This highlights the careful consideration required when implementing such a system.

Others have suggested implementing systems which respond directly to the person, for instance, by issuing a pre-recorded message or triggering lighting. While such systems may help to increase perceptions of surveillance or indeed improve visibility for human-led intervention, the evidence surrounding their effectiveness as a standalone measure for suicide prevention specifically is limited. In some instances, public awareness of such systems may also draw unnecessary attention to the site as a 'high-risk location'. Notably, a system installed on a bridge in Korea in 2012, which lit up 'messages of hope' as pedestrians walked by, was subsequently removed after an increase in deaths which seems to have followed the attention the project received. 

Additionally, AI could be used to identify potential patterns of behaviour preceding suicide. If possible, this could then help to enable earlier rescue responses. For example, in Seoul they are trialling the use of AI to learn behaviours that may signal distress in order to then predict risk and alert a rescue team where needed. A similar system is being trialled on the metro in Montreal, with claims that it can accurately predict one in four people at risk of suicide. If a suicide can be 'predicted' in this way, it may be that approaching trains can be slowed down and/or the appropriate people are alerted in time. 

Challenges of smart surveillance technology

Whilst there are documented advantages to the use of smart surveillance technology for suicide prevention, challenges to using these technologies have also been reported. These can relate to practical barriers of installation, such as funding the technologies – which can be very costly to procure and then complicated to put in place and maintain. This may be particularly true of rural locations with less infrastructure to support the use of these technologies. The ongoing monitoring of technology can also be complex to orchestrate in terms of finding appropriate staff, ensuring they are trained to operate the technology, and that there are resources in place to protect their welfare as best as possible.

As an example, the devastating impact of railway suicides on train drivers has been widely documented. First-hand accounts have highlighted that no amount of training can prepare train drivers for this tragic reality, but that having psychological and organisational support in place can help them to manage the difficult consequences they may face such as reliving the event, difficulty sleeping and possibly post-traumatic stress disorder. Finally, one of the biggest challenges is ensuring there is an appropriate, timely and sufficient response in place so that an incident can be identified and prevented, and the individual supported and kept safe. 

We must also consider how the wider public feel about being under surveillance. With phrases like 'Nanny state' and 'Big Brother' being used to show contempt for what is perceived as excessive surveillance and government interference, it must be acknowledged that technologies such as these are not always welcome. Some people have fears regarding data privacy, and with notable examples of CCTV images of a suicide attempt being distributed to the public, as well as cyber-attacks resulting in sensitive mental health records being leaked online, these concerns may not be unfounded. Managing such sensitive data is a tremendous responsibility and it is essential that the public can trust that it will be handled appropriately. 

The use of both surveillance and AI can also present issues and concerns around fairness and equality. Black communities in particular may have greater mistrust of surveillance due to institutional racism, for example research into public attitudes towards CCTV identified worries about young Black men being disproportionately targeted by surveillance technology. There are also examples being documented of AI being 'trained' on data which under-or mis-represents specific races or genders. In practice, this could mean that facial recognition for example is less accurate at detecting minoritised groups or those who are traditionally underrepresented. Similarly, oversights in the development of systems could lead some vulnerable individuals to be 'missed' by certain technologies, for example if the surveillance technology triggers an audio-based alert to warn someone they are in a dangerous area, this may not be accessible to people with hearing impairments. 

Clearly, there are many factors to consider when contemplating the use of surveillance technologies, and research is needed to understand these in greater depth. The work we are doing aims to explore these factors in real world contexts.

Questioning, understanding and evaluating suicide surveillance technologies (QUESST)

For the past year we have been conducting a project funded by the National Institute of Health and Care Research (NIHR), which is an evaluation of the use of smart surveillance technologies for suicide prevention in public places, to investigate whether and when they are effective (i.e., can they help reduce suicides and suicide attempts? Are there any potential (unintended) harms? What are the possible barriers to their effectiveness?). We are also examining the cost-effectiveness of these technologies (i.e., are they financially viable over time and how does their cost compare to having other measures in place, for example full time staff on location). Moreover, a process evaluation at three 'high-risk' sites across the UK, (including a bridge, a clifftop and a railway location) is using ethnographic and qualitative data to explore the perspectives of key stakeholders who may be impacted by these technologies (including those with lived experience through interviews and online ethnography) or those who are responsible for commissioning, implementing and/or monitoring the technologies. 

This work, and the results of our systematic review and national surveys of professional stakeholders working in the field, will inform guidance that can be distributed to those using or considering implementing surveillance technology for suicide prevention in public places. We are also aiming to share key principles from this work to help shape national policy around suicide prevention. 

Given the far-reaching aims of our project, this calls for work beyond the expertise of psychology alone.

Interdisciplinarity and knowledge exchange

Preventing suicides in public spaces can be complex, and so research in this area can require integrated and multifaceted approaches. As such, work in this area can be incredibly disparate and needs to look beyond psychology, psychiatry and public health. Indeed, even within academic journals research can span disciplines such as ergonomics, engineering, transport, health and safety, and computer science. Due to the applied nature of the work, relevant information is also often found in policy documents, presentations, and government or industry reports. While this poses barriers to collating information and ensuring it is accessible to the whole team and beyond, it also illustrates the importance of looking to collaborate with those outside of our field to most effectively address real-world issues such as this.  

Fortunately, in addition to those with expertise in suicide prevention working on the project, our work benefits from a multi-disciplinary team that includes experts in public health evaluation, health economics, ethnography, cyberpsychology, and the law.  We also work closely with Samaritans, our industry partners working in suicide prevention in public places, and a steering group of individuals with lived experience of suicide or bereavement by suicide. These stakeholder relationships are particularly important as another challenge is this being 'real world research', in that it's not controlled lab experiments, but a tragic reality where people are still dying by suicide contemporaneously to us doing this work. Collaborating with those listed above allows insights into different aspects of this reality to help us piece together this complex and multi-faceted puzzle.

Learning from lived experience

Even as psychologists who have studied it for several years, the topic of suicide doesn't get any easier. It can often be overwhelming, but it is nevertheless a privilege for us to have people  share their experiences with us. Given the need to 'change the narrative' and 'start the conversation' around suicide, hearing, learning from and sharing the voices of those who have lived and living experience of suicidality and/or bereavement by suicide is a crucial step. The National Suicide Prevention Alliance advisory group are a core part of our team and their expertise and input has been instrumental at all stages of the project in helping us to conduct this research appropriately, sensitively, and in a way that is most likely to be of benefit to those who need it. The qualitative work also highlights the experiences of those impacted by surveillance technologies, ensuring those experiences remain central to how effectiveness is understood beyond the statistics. 

We will have some findings from this project to share soon, and we hope that they will be of use to the many individuals who work tirelessly to help prevent suicide and improve the lives of those struggling with suicidality. 

It feels important to conclude with a reminder that, beyond preventing suicides at the point of crisis, there are many opportunities to intervene before someone reaches that place. There is great work being done to help better understand suicide and how to improve the conditions or factors that may lead someone to consider it. An essential part of this is starting conversations about suicide and changing the narrative so that it is a topic met with empathy, compassion, openness, and not silence. 

Disclaimer

This study/project is funded by the National Institute of Health and Care Research [Public Health Research (NIHR151521)]. The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.

Dr Bethany Cliffe, Research Fellow, University of Westminster

Dr Laura Joyner, Research Fellow, University of Middlesex 

Dr Lisa Marzano, Professor, University of Middlesex

Dr Jay-Marie Mackenzie, Reader, University of Westminster

Dr Ian Marsh, Reader, Canterbury Christ Church University 

Acknowledgements

Thank you to the group of NSPA (National Suicide Prevention Alliance) Lived Experience Influencers who contribute across this project. Thank you also to the wider project team (Professor Peter Craig, Alex Dark, Dr Manuela Deidda, Dr Carlisle George, Professor Keith Hawton, Amina Mahmood, Dr Robin Pharoah, Anthony Purvis, Ahmed Abdelsabour, and Dr Philip Worrall) and steering group (Professor Rachel Aldred, Professor Amy Chandler, Professor Frank De Vocht, Fiona Malpass, Professor Keith Hawton, Dr Emma McIntosh and Dr Hilary Norman).

If you have been affected by any of the topics covered in this article, please contact the Samaritans anytime for free on 116 123, or you can email them at [email protected]